Verda review

Facts from Verda’s own pricing and docs pages, September 202684 read from a quoted sentence on the page we archived.

Summary

Verda rents GPUs from $0.18 per GPU-hour on demand, bills 10-minute prepaid increments, and asks no minimum commitment. Verda does not publish an egress rate.

SOC 2 Type IIISO 27001GDPR · DPANo training on your dataSpend limits
From
$0.18/GPU-hr
Billing
10-minute prepaid increments
Minimum commitment
None
Egress
not stated
Uptime SLA
No SLA
Registered in
🇫🇮
Visit Verda
On-demand pricing
×1, on demand · RTX PRO 6000 96GB
$1.91/hr
Deploy
×4, on demand · H100 SXM5 80GB
$3.42/hr
Deploy
×1, on demand · H200 SXM5 141GB · 44 vCPU · 170 GB RAM
$4.50/hr
Deploy
×8, on demand · L40S 48GB
$1.50/hr
Deploy
×2, on demand · B200 SXM6 180GB
$6.68/hr
Deploy

See all 136 Verda offers →

What’s good
Lowest published on-demand rate is $0.18/GPU-hr (NVIDIA Tesla V100 16GB), with no minimum commitment
Billing is 10-minute prepaid increments
Self-serve signup — no sales call to get started
Free credits available: credit coupons at promotions and hackathons
Commitment discounts: 2% to 25% for 1 month to 2 years
Spend controls: low-balance alerts (default <48 h runway)
Plans
On-demand instancesSpot instancesReserved (long-term)Instant clusters
From$0.17 (V100)

Verda · On-demand instances

Every on-demand rate we hold for Verda, per GPU per hour, cheapest first.

×2, on-demand · Tesla V100 16GB
$0.18/hr
Deploy
×2, on-demand · RTX A6000 48GB
$0.62/hr
Deploy
×1, on-demand · RTX 6000 Ada 48GB · 10 vCPU · 60 GB RAM
$1.13/hr
Deploy
×1, on-demand · A100 SXM4 40GB · 22 vCPU · 120 GB RAM
$1.28/hr
Deploy
×8, on-demand · L40S 48GB
$1.50/hr
Deploy
×1, on-demand · A100 SXM4 80GB · 22 vCPU · 120 GB RAM
$1.74/hr
Deploy
×1, on-demand · RTX PRO 6000 96GB
$1.91/hr
Deploy
×4, on-demand · H100 SXM5 80GB
$3.42/hr
Deploy
×1, on-demand · H200 SXM5 141GB · 44 vCPU · 170 GB RAM
$4.50/hr
Deploy
×2, on-demand · B200 SXM6 180GB
$6.68/hr
Deploy
×2, on-demand · B300 SXM6 268GB
$8.12/hr
Deploy
×2, on-demand · GB300 SXM6 288GB
$9.34/hr
Deploy

Prices as last read from Verda’s own pages. Open a card for every provider renting it.

About half of on-demand: $1.64 (H100)

Verda · Spot instances

Every spot rate we hold for Verda, per GPU per hour, cheapest first.

×2, spot · Tesla V100 16GB
$0.09/hr
Deploy
×2, spot · RTX A6000 48GB
$0.31/hr
Deploy
×2, spot · RTX 6000 Ada 48GB
$0.57/hr
Deploy
×8, spot · A100 SXM4 40GB
$0.64/hr
Deploy
×2, spot · L40S 48GB
$0.75/hr
Deploy
×2, spot · A100 SXM4 80GB
$0.87/hr
Deploy
×2, spot · RTX PRO 6000 96GB
$0.95/hr
Deploy
×8, spot · H100 SXM5 80GB
$1.71/hr
Deploy
×1, spot · H200 SXM5 141GB · 44 vCPU · 170 GB RAM
$2.25/hr
Deploy
×4, spot · B200 SXM6 180GB
$3.34/hr
Deploy
×4, spot · B300 SXM6 268GB
$4.06/hr
Deploy
×2, spot · GB300 SXM6 288GB
$4.67/hr
Deploy

Prices as last read from Verda’s own pages. Open a card for every provider renting it.

2% (1 mo)

Verda · Reserved (long-term)

The rest of what Verda lists for this plan, as its own page states it:

3% (3 mo) · 4% (6 mo) · 8% (1 yr) · 25% (2 yr) off on-demand

$4.28 (H200)

Verda · Instant clusters

Every on-demand rate we hold for Verda, per GPU per hour, cheapest first.

×2, on-demand · Tesla V100 16GB
$0.18/hr
Deploy
×2, on-demand · RTX A6000 48GB
$0.62/hr
Deploy
×1, on-demand · RTX 6000 Ada 48GB · 10 vCPU · 60 GB RAM
$1.13/hr
Deploy
×1, on-demand · A100 SXM4 40GB · 22 vCPU · 120 GB RAM
$1.28/hr
Deploy
×8, on-demand · L40S 48GB
$1.50/hr
Deploy
×1, on-demand · A100 SXM4 80GB · 22 vCPU · 120 GB RAM
$1.74/hr
Deploy
×1, on-demand · RTX PRO 6000 96GB
$1.91/hr
Deploy
×4, on-demand · H100 SXM5 80GB
$3.42/hr
Deploy
×1, on-demand · H200 SXM5 141GB · 44 vCPU · 170 GB RAM
$4.50/hr
Deploy
×2, on-demand · B200 SXM6 180GB
$6.68/hr
Deploy
×2, on-demand · B300 SXM6 268GB
$8.12/hr
Deploy
×2, on-demand · GB300 SXM6 288GB
$9.34/hr
Deploy

Prices as last read from Verda’s own pages. Open a card for every provider renting it.

Minimum termNoneNone1 month to 2 yearsNone on-demand; reserved from 1 month
BillingPrepaid in 10-minute increments; unused portion refundedPrepaid, 10-minute incrementsUp-front for the contract lengthPay-as-you-go per GPU, no minimums
Uptime SLANo SLA; availability not warrantedNone; evictable at any pointNo SLA; availability not warrantedNo SLA; availability not warranted
TenancyVMs, 1×/2×/4×/8× GPUs, NVLink modelsSame instances, evictable without warningSelf-service instances and clusters, paid up-frontMulti-node InfiniBand clusters up to 144 GPUs self-serve
ProvisioningAs little as 30 secondsSelf-serve, subject to spot GPU quotaVia console; custom clusters via salesUnder 20 minutes
PaymentCard; SEPA or invoice bank transfer (min €/$1,000)Card or bank transferCard or bank transfer; custom invoice for bare-metal clustersCard or bank transfer
Details
Billing & access
How you pay
Prepaid balance (one-time or auto top-up); long-term paid up-front
Currency
USD prices; EUR or USD top-ups; VAT added
Commitment discounts
2% to 25% for 1 month to 2 years
Free credits
Credit coupons at promotions and hackathons
Refunds
Fees non-refundable, except unused prepaid increments on early termination
When funds run out
Instances stopped and volumes deleted at zero balance; restorable for 96 h
Late payment
9% annual interest; suspension
Provisioning guarantee
Account quotas cap concurrent GPUs; increases on request
What you get
GPUs listed
GB300 SXM6 288GB, B300 SXM6 268GB, B200 SXM6 180GB, H200 SXM5, H100 SXM5, A100 SXM4 80GB, A100 SXM4 40GB, Tesla V100, RTX PRO 6000, RTX PRO 6000 CC, L40S, RTX 6000 Ada, RTX A6000
Regions
FIN-01, FIN-02, FIN-03 — Helsinki, Finland
more
Docs also say pricing varies by location, so other locations may exist; only the Finnish sites are listed.
Node configuration
Per GPU: H100 30 CPU / 120 GB; H200 44 / 170 GB; B200 30 / 170 GB; GB300 32 / 225 GB
Interconnect
NVLink instances; InfiniBand clusters; GB300 NVL72
Datacenter tier
ISO 27001 certified data centres; PUE 1.2–1.3
Local storage
NVMe block volumes $0.20/GiB-month, 100k IOPS, 4 GB/s
Persistent storage
Block $0.20 · shared filesystem $0.20 · container registry $0.20 per GiB-month
Images
Ubuntu and JupyterLab images; startup scripts and reusable base images
Access methods
Console, API, CLI, SDK, Terraform, SkyPilot, SSH, JupyterLab
Model deployment
Serverless containers (vLLM, TGI, SGLang tutorials); dstack and SkyPilot integrations
Managed inference
Serverless containers for inference and batch
Managed Kubernetes
Pre-configured Kubernetes with Kueue, or Slinky (Slurm)
Serverless
Serverless containers, scale-to-zero; from $1.57 (L40S) to $8.50 (B300)
Kubernetes integration
Kubernetes with Kueue; Slinky for Slurm
Terraform
Terraform and OpenTofu providers
FAQ
How much does Verda cost?

As of September 2026, the lowest published on-demand rate is $0.18 per GPU-hour for the NVIDIA Tesla V100 16GB. Verda publishes an on-demand rate for 12 GPU models. Billing is 10-minute prepaid increments.

Does Verda have an uptime SLA?

Verda states No SLA; availability not warranted.

Is there a minimum commitment on Verda?

No. Verda states none; prepaid balance.

Does Verda charge for egress?

Verda does not publish an egress rate, so the cost of moving data out is not something a reader can check before signing up.

Can I sign up to Verda without approval?

Verda states that signup is self-serve.

Are Verda payments refundable?

Verda states: fees non-refundable, except unused prepaid increments on early termination.

Where are Verda's data centres?

Verda lists fIN-01, FIN-02, FIN-03 — Helsinki, Finland.

Who do you contract with at Verda?

The contracting entity is DataCrunch Oy.

Reviewed September 2026, from Verda’s own site. 84 facts, each quoted from the page it was read on.